Use of Capture-Recapture Analysis to Assess Reporting Completeness of Births to Hepatitis B Positive Women in New York City, 2013-2014

BACKGROUND: Mandatory prenatal screening and reporting requirements of hepatitis B virus (HBV) enables the New York City (NYC) Department of Health and Mental Hygiene (DOHMH) to identify and manage cases of births to HBV-positive women residing in NYC. Case identification is critical for the NYC Perinatal Hepatitis B Prevention Program (PHBPP) to effectively prevent perinatal HBV transmission. We sought to evaluate the completeness of the PHBPP surveillance system.

METHODS: We used a capture-recapture analysis that involved two separate data matches. Mothers of infants born in NYC from 5/1/2013-5/1/2014 identified through the NYC immunization information system (IIS) were matched to persons with HBV-positive lab reports in the NYC Electronic Clinical Laboratory Reporting System (ECLRS) with a collection date between 8/1/2012-7/31/2014. A multi-level deterministic match was conducted, incorporating mother name, date of birth, phone number, and address. Infants born to HBV-positive women identified through this process were matched against the NYC PHBPP database, which contains all births to HBV infected women reported to the DOHMH, using a similar matching algorithm.  Non-matching records were investigated to confirm NYC residency and HBV status at delivery; Petersen’s estimator was used to estimate the number of cases not captured by the existing PHBPP surveillance system.

RESULTS: The ECLRS/IIS match identified 1,685 infants born to HBV-positive women; 1,516 were found in the PHBPP database, of which 1,493 were confirmed cases and 23 were not true cases or resided outside of NYC at delivery. The remaining 169 infants were not found in the PHBPP database and required investigation; 55 were confirmed to have been born to HBV-positive women residing in NYC at delivery. A total of 1,554 cases were reported to the PHBPP; 61 (3.9%) were not found in the ECLRS/IIS match. Petersen’s method estimated 1,611 cases, indicating that 55 (3.4%) were not identified by the existing PHBPP surveillance system, and 2 (0.1%) were not identified by either the ECLRS/IIS match or the PHBPP surveillance system. The ECLRS/IIS match resulted in a 3.5% increase in case identification. The existing PHBPP surveillance system was estimated to capture 96.5% of cases.  

CONCLUSIONS: The results indicate that the PHBPP surveillance system captures a very high percentage of estimated cases. The ECLRS/IIS match increased case ascertainment; regular use of the method could improve future case identification efforts. However, the ECLRS/IIS match did not find all cases reported to the PHBPP, suggesting that the matching algorithm could be improved.